Always treat both 0 and 1 as root prim | Voters

Always treat both 0 and 1 as root prim

Tapple Gao

In 1-prim objects, root prim has link number 0. In other prims, root prim has link number 1 (

=LINK_ROOT

). This behavior should change:

It's a pain to make scripts work with both types of objects
Link numbers are
so close
to already using 1-based indexing, like everything else in SLua has been already changed to. This change would complete that transition

For functions that accept a link number:

0
and
1
should both be treated equally as the root prim number, regardless of prim count
This change should be made in SLua at least, but possibly in LSL and Mono as well. (it's probably easier to change both)
I can't think of a reason this would break existing LSL content, as it only makes the functions do something useful where they previously did nothing (Suzanna thought of a theoretical reason it could break code below)

For functions that return a link number:

In LSL, they must be left alone, as this check is somewhat common to test for 1-prim vs multi-prim:
if (llGetLinkNumber())
In SLua, they should be changed to always return 1 (
=LINK_ROOT
). This makes them more consistent, and SLua has no backwards compatibility concerns yet (This is the main reason this proposal is in SLua-Alpha and not Scripting Bugs)

These functions accept link numbers, and return a prim property:

ll.AvatarOnLinkSitTarget
ll.GetLinkKey
ll.GetLinkMedia
ll.GetLinkName
ll.GetLinkNumberOfSides
ll.GetLinkPrimitiveParams
ll.GetLinkSitFlags
ll.GetObjectLinkKey
ll.IsLinkGLTFMaterial

These functions accept link numbers, and modify them:

ll.BreakLink
ll.ClearLinkMedia
ll.LinkAdjustSoundVolume
ll.LinkParticleSystem
ll.LinkPlaySound
ll.LinkSetSoundQueueing
ll.LinkSetSoundRadius
ll.LinkSitTarget
ll.LinkSetSoundRadius
ll.MessageLinked
ll.SetLinkAlpha
ll.SetLinkCamera
ll.SetLinkColor
ll.SetLinkGLTFOverrides
ll.SetLinkMedia
ll.SetLinkPrimitiveParams
ll.SetLinkPrimitiveParams({PRIM_LINK_TARGET})
ll.SetLinkPrimitiveParamsFast
ll.SetLinkPrimitiveParamsFast({PRIM_LINK_TARGET})
ll.SetLinkRenderMaterial
ll.SetLinkSitFlags
ll.SetLinkTexture
ll.SetLinkTextureAnim
ll.SetPrimitiveParams({PRIM_LINK_TARGET})
ll.SitOnLink

These functions return a link number:

LLEvents.link_message
ll.CastRay
ll.DetectedLinkNumber
DetectedEvent.getLinkNumber
ll.GetLinkNumber
ll.GetObjectDetails({OBJECT_LINK_NUMBER})

March 3, 2026

Tapple Gao

We tried making the smallest Lua code (by compiled size) to iterate over all valid link numbers. Here's the ones I thought were the most interesting:

Smallest bytecode, by Suzanna:

local numPrims = ll.GetNumberOfPrims()
local isUnlinked = numPrims == 1 and 0
for primId = isUnlinked or 1, isUnlinked or numPrims do
    print(primId, ll.GetLinkName(primId))
end
print(ll.GetUsedMemory() - 2267) -- > 157

Smallest readable bytecode, by Kristy:

local numPrims = ll.GetNumberOfPrims()
local startPrim = 1
if numPrims == 1 then
    numPrims = 0
    startPrim = 0
end
for primId = startPrim, numPrims do
    print(primId, ll.GetLinkName(primId))
end
print(ll.GetUsedMemory() - 2267) -- > 162

Smallest reusable iterator, by Tapple and Suzanna:

local function allPrims()
    local numPrims = ll.GetNumberOfPrims()
    local isUnlinked = numPrims == 1 and 0
    return function(last, i)
        if i < last then return i + 1 end
    end, isUnlinked or numPrims, isUnlinked and -1 or 0
end
for primId in allPrims() do
    print(primId, ll.GetLinkName(primId))
end
print(ll.GetUsedMemory() - 2267) -- > 312

Tapple Gao

If this bug were fixed, these could just be:

for primId = 1, ll.GetNumberOfPrims() do
    print(primId, ll.GetLinkName(primId))
end
print(ll.GetUsedMemory() - 2267) -- > 128

Journey Bunny

integer rootLinkNum = !!llGetLinkNumber();

Now rootLinkNum is correct number.

SuzannaLinn Resident

Journey Bunny

In SLua:

local rootLinkNum = if ll.GetLinkNumber() == 0 then 0 else 1

Tapple Gao

SuzannaLinn Resident the smallest Lua code (by text size) to get the root prim I can think of is

local rootLinkNum = math.sign(ll.GetLinkNumber())

Suzanna's is smaller by bytecode size (72 bytes vs 79)

SuzannaLinn Resident

I like the idea.

I would do it only in SLua in the

library (1-based LL functions).

I would leave untouched the

llcompat

library (0-based LL functions) and LSL.

Probably there are things like this in LSL...

-- orange box that turns blue when someone sits
local function setColor()
    llcompat.SetLinkColor(0, vector(1, 0.5, 0), ALL_SIDES)
    llcompat.SetLinkColor(1, vector(0, 0.5, 1), ALL_SIDES)
end
LLEvents:on("changed", function(change)
    if bit32.btest(change, CHANGED_LINK) then
        setColor()
    end
end)
setColor()

Lou Netizen

Modifying functions that return a link number to always return LINK_ROOT in single-prim objects will break a good deal of existing scripting. Testing if a link number is zero is a very common way to test if a script is in a single-prim object.

Tapple Gao

Lou Netizen Added a note to only consider doing that in SLua